Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/conorbronsdon/agent-memory-kit/endgit clone --depth 1 https://github.com/conorbronsdon/agent-memory-kitWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00014 | $0.01412 |
| Opus 5 | $0.00007 | $0.00706 |
| Sonnet 5 | $0.00003 | $0.00282 |
| Haiku 4.5 | $0.00001 | $0.00141 |
Grade A, and why
end scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 122 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/end — Close Session
Capture the session: write the episodic log, update active state, and propose 0–2 durable memories. Capture only — deep curation belongs to /dream.
1. Auto-extract the session summary
Scan the full conversation top to bottom and extract automatically — don't wait to be asked.
Scope guard (optional but recommended). If you keep separate personal and work contexts, filter here. Log only the items that belong to this context. When in doubt, leave it out — a missing entry is fixable, a leaked entry causes drift.
Extract:
- Topics covered — what was worked on (tool calls, file edits, discussion).
- Decisions made — anything concluded or chosen, with the rationale, not just the choice.
- Rejected alternatives — for each real decision, what else was considered and why it lost. If a bug was fixed, the wrong theory tried first. This "failed hypothesis" record is what stops a future session from repeating the same wrong starting point.
- State changes — priorities that shifted, threads that opened or closed, blockers that resolved.
- Open threads — unfinished items or things waiting on someone else.
- Next actions — what needs to happen before or at the next session.
Present the summary for a quick confirmation before writing. Goal: a fresh session tomorrow should read state/current.md alone and know exactly where things stand.
2. Write the session log
Run date +%Y-%m-%d for TODAY and date +%H:%M for TIME. Append to sessions/{TODAY}.md (create with a # Session Log: {TODAY} header if new):
## Session: {TIME}
### Topics
- {topic}
### Decisions
- {decision}
### Open Threads
- {thread}
### Next Actions
- {action}
3. Update state
Always update state/current.md:
- Add new open threads; remove completed items.
- Adjust priorities if anything shifted.
- Mark touched items with
*(updated M/D)*. - Update the "Last updated" date.
state/current.md is for context and attention, not scheduling. Keep volatile schedule data (dates, numbers that change) in its source of truth and reference it, don't copy it here.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 122 lines · 14 tokens per session scan A 95f8d0033ee7
end is a command published in the GitHub repository conorbronsdon/agent-memory-kit (2 stars, last pushed 28d ago), licensed MIT. It adds 14 tokens to every session and 1,412 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
release
Release a new version: bump version, update changelog, tag, push, and create a GitHub release.
benchmark
Run a TAU-bench evaluation end-to-end and present results.
checkout-branch
Switch to an existing branch by checking out its worktree, or creating one if needed.
kayba-agent-instructions
The kayba CLI interacts with the Kayba hosted API (https://use.kayba.ai). Auth: set KAYBAAPIKEY env var or pass --api-key to every command.
create-branch
Create a new git branch with an associated worktree following the project naming convention.
create-pr
Create a pull request for the current branch against main.